ASCO/SSO Review of Current Role of Risk-Reducing Surgery in Common Hereditary Cancer Syndromes

Transcription

1 V O L U M E 2 4 N U M B E R 2 8 O C T O B E R JOURNAL OF CLINICAL ONCOLOGY R E V I E W A R T I C L E From the Departments of Surgery and Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY; Department of Surgery, Emory University School of Medicine, Atlanta, GA; Department of Surgery, Washington University School of Medicine, St Louis, MO; Department of Obstetrics and Gynecology, Duke University Medical Center, Durham, NC; Department of Obstetrics and Gynecology, Cedars-Sinai Medical Center and David Geffen School of Medicine at the University of California Los Angeles, Los Angeles, CA; Department of Obstetrics and Gynecology, Washington University School of Medicine, St Louis, MO; Department of Surgery, The University of Texas M.D. Anderson Cancer Center, Houston, TX; Department of Clinical Cancer Genetics, City of Hope National Medical Center, Duarte, CA; Department of Surgery, Fox Chase Cancer Center; Departments of Medicine and Genetics, University of Pennsylvania School of Medicine, Philadelphia, PA; Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD; Department of Colorectal Surgery, The Cleveland Clinic Foundation, Cleveland, OH; and the Department of Medicine, The University of Michigan Medical Center, Ann Arbor, MI. Submitted October 7, 2005; accepted June 30, Authors disclosures of potential conflicts of interest and author contributions are found at the end of this article. Address reprint requests to José G. Guillem, MD, MPH, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, Room C-1077, New York, NY 10021; guillemj@mskcc.org. ASCO/SSO Review of Current Role of Risk-Reducing Surgery in Common Hereditary Cancer Syndromes José G. Guillem, William C. Wood, Jeffrey F. Moley, Andrew Berchuck, Beth Y. Karlan, David G. Mutch, Robert F. Gagel, Jeffrey Weitzel, Monica Morrow, Barbara L. Weber, Francis Giardiello, Miguel A. Rodriguez-Bigas, James Church, Stephen Gruber, and Kenneth Offit A B S T R A C T Although the etiology of solid cancers is multifactorial, with environmental and genetic factors playing a variable role, a significant portion of the burden of cancer is accounted for by a heritable component. Increasingly, the heritable component of cancer predispositions has been linked to mutations in specific genes, and clinical interventions have been formulated for mutation carriers within affected families. The primary interventions for mutations carriers for highly penetrant syndromes such as multiple endocrine neoplasias, familial adenomatous polyposis, hereditary nonpolyposis colon cancer, and hereditary breast and ovarian cancer syndromes are primarily surgical. For that reason, the American Society of Clinical Oncology (ASCO) and the Society of Surgical Oncology (SSO) have undertaken an educational effort within the oncology community. A joint ASCO/SSO Task Force was charged with presenting an educational symposium on the surgical management of hereditary cancer syndromes at the annual ASCO and SSO meetings, resulting in an educational position article on this topic. Both the content of the symposium and the article were developed as a consensus statement by the Task Force, with the intent of summarizing the current standard of care. This article is divided into four sections addressing breast, colorectal, ovarian and endometrial cancers, and multiple endocrine neoplasia. For each, a brief introduction on the genetics and natural history of the disease is provided, followed by a detailed description of modern surgical approaches, including a description of the clinical and genetic indications and timing of prophylactic surgery, and the efficacy of prophylactic surgery when known. Although a number of recent reviews have addressed the role of genetic testing for cancer susceptibility, including the richly illustrated Cancer Genetics and Cancer Predisposition Testing curriculum by the ASCO Cancer Genetics Working Group (available through this article focuses on the issues surrounding the why, how, and when of surgical prophylaxis for inherited forms of cancer. This is a complex process, which requires a clear understanding of the natural history of the disease and variance of penetrance, a realistic appreciation of the potential benefit and risk of a risk-reducing procedure in a potentially otherwise healthy individual, the long-term sequelae of such surgical intervention, as well as the individual patient and family s perception of surgical risk and anticipated benefit. J Clin Oncol 24: This article has been published jointly by invitation and consent in both the Journal of Clinical Oncology and the Annals of Surgical Oncology Society of Surgical Oncology. All rights reserved. No part of this document may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without written permission by the Society of Surgical Oncology This article has been published jointly by invitation and consent in both the Journal of Clinical Oncology and the Annals of Surgical Oncology Society of Surgical Oncology. All rights reserved. No part of this document may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without written permission by the Society of Surgical Oncology X/06/ /$20.00 DOI: /JCO HEREDITARY BREAST CANCER In 1865 a French surgeon Paul Broca described his wife s pedigree of four generations of breast cancer. 1 This observation was cited a century later when Henry Lynch et al 2 described 34 families with two or more first-degree relatives with breast cancer, including a description of hereditary breast and ovarian cancer. In 1990, Hall noted a linkage between chromosome 17q and early-onset breast cancer. 3 Narod demonstrated a linkage to this same site in the hereditary breast and ovarian cancer syndrome. 4 The site was soon cloned. 5 In 1994, a second breast cancer susceptibility gene, BRCA2, was linked to chromosome 13q. 6 The intervening years have seen an explosion of information regarding the genetic repressor mechanism by which mutations in BRCA1 and BRCA2 genes cause inherited susceptibility to breast cancer. The Li-Fraumeni syndrome, associated with breast cancer and soft tissue sarcomas, was first described as a kindred in ; this is caused by a germline 4642

2 Current Role of Surgery in Hereditary Cancer mutation in the p53 gene. In 1997, germline mutations of the PTEN gene were shown to be associated with Cowden disease, an inherited breast and thyroid cancer syndrome also known as the multiple hamartomasyndrome. 8,9 Amutationonchromosome11q,associated with ataxia telangiectasia, has been shown to convey a three- or fourfold increased risk of a variety of cancers. Female carriers of the mutated AT gene (ATM) are at approximately five times the risk of the general population for developing breast cancer. 10 It was initially assumed that a host of other major genetic mutations would be found to account for all familial breast cancer. This hope has not yet been realized. The primary genes associated with risk for breast cancer are BRCA1 and BRCA2. However, it is now possible to define the relative risk of having such a mutation from a detailed family history to guide recommendations for genetic testing (Table 1) When a three- or four-generation family tree suggests sufficient risk, testing for these genetic mutations is available commercially, and is often covered at least partially by insurance. The concern that insurance or employment discrimination might result from a positive test has not materialized. However, it should be noted that the perception of potential discrimination continues to pose a significant problem for the high-risk patient. This may explain, in part, why genetic testing for women whose family history suggests a significant risk is still not widely utilized. Indications for Risk-Reducing Surgery Risk-reducing surgery for breast cancer is indicated for three different populations of women. The first group comprises those women who have undergone genetic testing and have been found to express a mutated gene associated with high penetrance breast cancer. Although the risk is specific to a population rather than the individual, this group has a dramatically increased risk of developing breast cancer between 56% and 87% in the case of BRCA1 and BRCA2 mutation carriers, respectively. The specific population risk varies somewhat, depending on the specific mutation. Not only is lifetime risk greatly increased, but these mutations carry a risk of breast cancer developing much earlier in life than is seen with sporadic breast cancer. The second population group comprises those women who present with a strong family history suggestive of hereditary breast cancer, but who test negative for BRCA1 or BRCA2. These individuals are clearly at increased risk, but it is impossible at present to define the magnitude of this increased risk with any accuracy. The third population group is composed of those women who have a strong family history but have not been tested. When a hereditary cancer syndrome is suspected in a family, the most important person to test is the relative affected with early breast or ovarian cancer. Once a mutation is identified in an affected family member, the test is considered informative in that family and suitable for testing at-risk individuals. If a woman has not inherited the mutation, she has only the same risk of developing breast cancer as women in the general population even a bit less, as the general population risk includes some women who are actually at increased risk. Consideration of prophylactic surgery in the absence of genetic testing should be strongly discouraged. However, patients who have tested negative, but are perceived to be obligate gene carriers based on a review of their pedigree may also be considered for prophylactic surgery. Surgical Considerations Prophylactic surgery for women at high risk of breast cancer has traditionally been discussed in terms of mastectomy. However, in virtually all animal models of breast cancer development, mastectomy has failed to eliminate the risk completely Clinical studies of reduction mammoplasty, however, show that this procedure although leaving a generous amount of breast tissue is associated with a decreased lifetime risk of breast cancer when compared with the sisters of this cohort; the decrease in risk approached 40%. 18 Large series of subcutaneous mastectomy (nipple sparing) were associated with few subsequent breast cancers. These series were not limited to women of defined risk, however. Dr Lynn Hartmann reviewed the Mayo Clinic series and attempted to define population risk from family history. Her analysis suggested a greater than 90% prevention of subsequent breast cancer by such mastectomies. 19 Although a randomized trial to address this subject in the purest fashion is not feasible, several studies have recently emerged comparing those mutation carriers who elected prophylactic bilateral mastectomy with thosewhodidnot. 20 Reassuringly,thisalsodemonstratedariskreduction of at least 90%. A larger series has now been published confirming a risk reduction of at least 90% from prophylactic mastectomy in mutation carriers. 21 Modern surgical approaches to surgical prophylaxis include total mastectomy without an axillary lymph node dissection, skin-sparing total mastectomy, and subcutaneous mastectomy under a new name, nipple-sparing mastectomy (Table 2). A final procedure being discussed is areolar-sparing mastectomy. This procedure is a skinsparing mastectomy in which the skin of the areola is also spared, and only the nipple and breast are removed. There is insufficient experience with this procedure to make any definitive statements about it, either in terms of its cosmetic advantages (if any) or its oncologic downside (if any). The case for subcutaneous mastectomy arises from the perceived cosmetic advantage of preserving the nipple and areola. To the degree that breast tissue is left in the nipple and immediately beneath the areola, however, the risk of cancer arising is clear. To the Table 2. Types of Prophylactic Mastectomy Table 1. Genetic Syndromes and Lifetime Breast Cancer Risk Syndrome Risk (%) BRCA BRCA Cowden Syndrome Li-Faumeni Syndrome Increased Heterozygous AT Increased Type of Mastectomy Total Skin sparing Subcutaneous (nipple sparing) Areola sparing Features All breast tissue, nipple, areola, skin of breast; most effective All breast tissue, nipple, areola; appears equally effective to total All breast tissue, except nipple, areola; anecdotally less effective Few performed, no follow-up data NOTE. Adapted from Srivastava et al. 11 Abbreviation: AT, ataxia telangiectasia. NOTE. All breast tissue means all apparent breast tissue; a tiny fraction always persists at the periphery of the mastectomy

3 Guillem et al degree that this breast tissue is removed, the cosmetic advantage of nipple and areolar sparing is lost, since what remains is essentially a full-thickness skin graft that may heal with considerable distortion. The ability of plastic surgeons to reconstruct and tattoo normalappearing nipples and areolae detract from any putative benefit. Pathologic series showing cancer cells and in situ cancer in the nipple ducts of breast specimens have led to admonitions against subcutaneous mastectomy from surgical oncologists. Although these warnings seem reasonable, they are supported only by anecdotal, not substantive trial data. In the Rebbeck study, 21 both failures were in subcutaneous mastectomy patients. As the entire purpose of bilateral prophylactic mastectomy is put at risk by the breast tissue left behind, a strong case can be made for bilateral total mastectomy. This may be performed with a skin-sparing technique. The preserved skin envelope is then immediately refilled by either a transabdominal myocutaneous (TRAM) flap or a latissimus flap. In the latter case, an underlying prosthesis or tissue expander of appropriate size may be used. Whenever a prosthesis is implanted, overlying muscle beneath the skin allows a more natural look and feel. The case for skin-sparing mastectomy is made on a cosmetic basis. It is supported by the finding of large series performed for cancer that show no suggestion of higher failure rates than in skin-ablating total mastectomy. 22 Again, these are not randomized, comparative studies that would allow a definitive statement to be made. The downside of this clear risk-reduction benefit is the total of loss of all nipple sensation, a loss that women often state is considerably greater than they had anticipated it would be, hindering sexual arousal. Satisfaction with this procedure exists for those who have replaced a high level of chronic anxiety about breast cancer with a desire for favorable cosmetic outcome. Prophylactic mastectomy should be considered with other prophylactic surgical options (Table 3), including prophylactic salpingooophorectomy. Many patients find this a preferable prevention compared with prophylactic mastectomy, because it is a hidden procedure. It is approximately 90% effective in reducing the risk of subsequent ovarian cancer. Because ovarian cancer is difficult to detect at its earliest stages, it has an appeal purely in terms of preventing ovarian cancer. The data from Kauf et al 23 and Rebbeck et al 21,24 all suggest that a 90% reduction of ovarian cancer risk is accompanied by a roughly 50% reduction in the risk of breast cancer. Because of the dual benefits, and the ease with which it can generally be performed laparoscopically, prophylactic salpingo-oophorectomy is a first topic of conversation with mutation carriers. Clearly, this must be placed within a context of timing, family planning, and risk development. Nonsurgical Alternatives All prophylactic surgery must be weighed against the alternatives. The first alternative is surveillance with intervention for cause; the Table 3. Management Options for High-Risk Women Option Prophylactic bilateral mastectomy Prophylactic salpingo-oophorectomy Chemoprevention (tamoxifen) Surveillance (CBE, BSE, mammography, MRI) Abbreviations: CBE, clinical breast examination; BSE, breast self-examination; MRI, magnetic resonance imaging. second is chemoprevention. The observation from clinical trials of adjuvant tamoxifen for breast cancer demonstrated reduction in the risk of contralateral breast cancer. After 5 years of tamoxifen, this reduction approached 50%. The National Surgical Adjuvant Breast and Bowel Project P1 trial demonstrated the safety of this approach to chemoprevention in women at even moderately increased risk. As data from the Rebbeck study 24 suggest a 50% reduction in risk of breast cancer following oophorectomy, it has been postulated that tamoxifen may reduce the risk of inherited breast cancer. Narod et al 25 demonstrated a 30% to 40% reduction in the risk of contralateral breast cancer in mutation carriers who received tamoxifen as adjuvant therapy for their first breast cancer. This is at almost the same rate as that seen in general cancer patients. Taken with all the epidemiologic data now published on breast cancer risk after prophylactic oophorectomy, it would appear that breast cancer risk in mutation carriers is significantly reduced by decreasing the influence of estrogen on breast tissue. The ultimate phenotype of the tumor does not appear to reflect the influence of estrogen on promotion initiation based on these growing experiences. This would suggest that both BRCA1 and BRCA2 mutation carriers would benefit from taking tamoxifen for risk reduction. If chemoprevention lacks prospective validation in the mutationpositive population, the effectiveness of surveillance for early detection is also still being defined. This is of great importance, because close surveillance remains the option chosen by a majority of high-risk women. The significant number of interval cancers seen in BRCA1 and BRCA2 mutation women under surveillance has encouraged several trials of magnetic resonance imaging for breast screening of this high-risk population. At present, the general consensus regarding magnetic resonance imaging (MRI) is that the greater the experience, the more accurate the screening. MRI is far more sensitive than mammography, but even less specific. However, when applied to a population with a much higher incidence of breast cancer, the relative loss of specificity may still be acceptable, especially in view of the greater sensitivity afforded by this modality. Morris demonstrated this in a large retrospective study of asymptomatic women with a variety of high risk factors. 26 Brekelmans et al 27 followed nearly 1,200 women with familial risk factors, and added MRI to the other screening studies for women with BRCA1 or BRCA2 mutations. MRI appeared to be more cost effective and of greater benefit than other imaging modalities. Kuhl et al screened 462 women with mammography, ultrasound, and MRI, in addition to clinical breast exam. MRI was the most sensitive (96%, compared with 43% for mammography and 47% for ultrasound). 28 In yet another trial by Kriege et al, 29 MRI showed a 71% sensitivity compared with 36% for mammography, with a specificity of 88% for MRI versus 95% for mammography. The concern regardingmriisthatitslowerspecificityleadstohigherrecallratesandmore biopsies. In a population of greatly increased risk this may be tolerable. For women already anxious about increased risk, however, being recalled for additional films, additional imaging studies, or needle aspirations may surely prove anxiety provoking. Screening women at increased risk with dense breast tissue represents a major challenge. Breast MRI appears to offer sufficient benefit such that some have advocated an annual MRI with annual mammogram alternatively at 6-month intervals to provide the most intense screening possible. In addition, we routinely recommend monthly breast self-examination (which is already recommended for women at average risk). Although it has been difficult to measure the 4644 JOURNAL OF CLINICAL ONCOLOGY

4 Current Role of Surgery in Hereditary Cancer relative advantages and disadvantages of this technique in detecting interval breast cancers, we have found that breast self-exam is empowering to many women with increased risk. The report on prophylactic mastectomy by Meijers-Heijboer 20 included a woman who had elected surveillance and presented 1 year from the time of first evaluation with a 4-cm node-positive tumor. Clearly, the effectiveness of surveillance in a high-risk population is influenced by the compliance of the patient, the radiologic density of her breast tissue, the nodularity or consistency of the breast tissue to palpation, and the availability of excellent breast imaging. Decisions regarding prophylactic bilateral mastectomy and bilateral salpingo-oophorectomy, and the technique to be used, should be made by considering risk over the next decade rather than lifetime risk. In the future, continued progress may improve our ability to prevent cancer. Evidence of benefit from tamoxifen and prophylactic oophorectomy must be shared with these women. High-risk breast clinics offer the opportunity for structured, careful evaluation over time, and consideration of both chemoprevention and risk-reducing surgical interventions by those with experience in technique and communication. FAMILIAL ADENOMATOUS POLYPOSIS AND HEREDITARY NONPOLYPOSIS COLORECTAL CANCER Introduction Inherited colorectal cancer primarily comprises two syndromes which predispose to disease by germline mutation transmitted in an autosomal dominant fashion. Familial adenomatous polyposis (FAP), which accounts for less than 1% of the annual colorectal cancer burden, is caused by mutations in the tumor-suppressor adenomatous polyposis coli (APC) gene and is characterized by the presence of 100 or more adenomatous polyps in the colorectum, nearly 100% penetrance, and an inevitable risk of colorectal cancer at the average age of 40 if prophylactic colectomy is not performed. 30,31 In the less severe form, attenuated FAP, patients can present with fewer than 100 colorectal adenomas, which tend to be proximally located. Recently reported are biallelic germline mutations in the base-excision-repair gene MYH, which may account for 7.5% of patients with a classical FAP phenotype who have no demonstrable APC mutation. 32 MYHassociated polyposis shows an autosomal recessive pattern of inheritance, and often presents as attenuated polyposis. 33 Hereditary nonpolyposis colorectal cancer (HNPCC), also commonly referred to as the Lynch syndrome, accounts for 2% to 3% of all colorectal cancer and is due to a germline mutation in one of the DNA mismatch repair (MMR) genes (hmlh1, hmsh2, hmsh6, and PMS2). 34,35 HNPCC is characterized by early age of onset colorectal cancer, a predominance (70%) of lesions proximal to the splenic flexure, an increased rate of metachronous colorectal tumors, and a unique spectrum of benign and malignant extracolonic tumors. 31 Lifetime risk of colorectal cancer in HNPCC patients is approximately 80%. 36,37 Microsatellite instability (MSI), reflecting a deficiency in DNA repair secondary to the mutation in the MMR genes, is a common feature of HNPCC-related tumors. 31 Differences in penetrance, phenotypic expression and certainty of disease development mandate distinctly different surgical approaches in FAP and HNPCC, including the type and timing of risk-reducing colon and rectal surgery. FAP Surveillance (based on genetic testing or annual flexible sigmoidoscopy) of at-risk family members should begin around puberty(10 to 12 years of age). At-risk individuals belonging to families with an attenuated FAP phenotype should undergo screening with a colonoscope due to the proclivity of colonic adenomas to be located proximally in this variant. 38 In families with a demonstratedapc mutation, informative genetic testing is possible with the protein truncation test, gene sequencing, or analysis for large deletions and rearrangements. These methods detect mutations in 90% to 95% of FAP pedigrees. 39 It is worth noting that deletion studies are increasingly part of the standard clinical panel performed by reference laboratories. Patients with either a positive genotype or adenomatous polyps on sigmoidoscopy should undergo full colonoscopy to establish the severity of polyposis. Timing of surgery depends to some degree on the extent of polyposis, as the risk of colorectal cancer development is partially dependent on colon and rectal polyp burden. 40 Patients with mild polyposis and corresponding lower cancer risk can undergo surgery in their midteens. Patients with severe polyposis, severe dysplasia, tubulovillous architecture, multiple adenomas larger than 5 mm in size, and symptoms (including bleeding, persistent diarrhea, anemia, failure to thrive, and psychosocial stress) should undergo risk-reducing colorectal surgery as soon as is practical after diagnosis. 36 However, in carefully selected, fully asymptomatic cases with small adenomas, yet with a strong family history for aggressive abdominal desmoid disease, consideration can be given to delaying prophylactic colectomy, as the risk of a desmoid-related complications may be greater than the risk of developing a colorectal cancer. The three current surgical options for patients with FAP are total proctocolectomy with permanent ileostomy (TPC), total colectomy with ileorectal anastomosis (IRA), and proctocolectomy with ileal pouch anal anastomosis (IPAA). The selection of the optimal procedure for an individual patient is based on several factors, including characteristics of the FAP syndrome within the patient and patient s family, differences in likely postoperative functional outcome, preoperative anal sphincter status, and patient preference TPC with permanent ileostomy, although rarely chosen as a primary procedure, is employed in patients with unacceptably poor baseline sphincter function, an invasive cancer involving the sphincters or levator complex, or for patients in whom an IPAA is not technically feasible (secondary to desmoid disease and foreshortening of the small bowel mesentery leading to the inability to bring the ileal pouch to the anus). However, TPC is occasionally chosen as a primary procedure by patients who perceive that their lifestyle would be compromised by the frequent bowel movements (5 to 6/d) sometimes associated with the IPAA procedure. In addition to the issues mentioned herein, the key in deciding between an IPAA and an IRA relates primarily to the risk of rectal cancerdevelopmentiftherectumisleftinsitu.theriskofrectalcancer after IRA may be as high as 4% to 8% at 10 years, and 26% to 32% after 25 years. 44,45 This risk may be overestimated, however, as most studies were completed before the widespread availability of IPAA. 46 Thus, patients and physicians may have chosen an IRA even in the setting of extensive rectal disease, as TPC with permanent ileostomy was the only other available option at that time. The magnitude of risk in an individual patient is, however, related to the overall extent of colorectal polyposis. IRA may be considered for patients with fewer than 1,000 colorectal polyps (including attenuated FAP) and fewer

5 Guillem et al than 20 rectal adenomas, as these patients have a relatively low risk of rectal cancer. 36,40,47-49 Patients with severe rectal ( 20 adenomas) or colonic ( 1,000 adenomas) polyposis, an adenoma larger than 3 cm or an adenoma with severe dysplasia, should ideally undergo proctectomy. 36,40,47,48 The risk of secondary rectal excision, due to uncontrollable rectal polyposis or rectal cancer, may be estimated by the specific location of the causative APC mutation. 45,49,50 In one study, patients with a mutation located between codons 1250 and 1464 had a 6.2-fold increased risk of rectal cancer compared with those with mutation before codon 1250 or after codon 1464 (mean number of rectal polyps 42 v 22, respectively). 45 Although the concept of using genotype-phenotype relations to help guide the management of a specific patient is appealing, it is important to recognize that variability of phenotypic expression even within members of the same family suggest that the current basis for choosing between an IRA and an IPAA should be primarily on clinical, rather than genetic, grounds. The risk of polyp and cancer development after primary surgery is not limited to patients undergoing IRA. In patients undergoing IPAA, the pouch-anal anastomosis can be hand sewn after complete anal mucosectomy, or stapled to the anus, thereby leaving in situ a 1 cm to 2 cm segment of mucosa called the anal transition zone. Neoplasia may occur at the site of ileal pouch anastomosis, and the frequency appears to be greater after stapled anastomosis (range, 28% to 31%) than after mucosectomy and hand-sewn anastomosis (range, 10% to 14%). 51,52 Function may be better, however, after stapled anastomosis. 52 In the case of neoplasia developing at the anal transition zone after a stapled anastomosis, transanal mucosectomy can often be performed, followed by advancement of the pouch to the dentate line. 53 Of additional concern is the development of adenomatous polyps in the ileal pouch, which occurs in 35% to 42% of patients by 7 to 10 years of follow-up Consequently, after either procedure, lifetime surveillance of the rectal remnant (after IRA) or the ileal pouch (after IPAA) is required. 36 Another important consideration in choosing between an IPAA and IRA is postoperative bowel function and associated quality of life. Some studies have associated IPAA with higher frequency of both daytime and nocturnal bowel movements, higher incidence of passive incontinence and incidental soiling, and greater postoperative morbidity However, other studies comparing IPAA and IRA have shown equivalent functional results 60 and quality of life. 61 Therefore, although the choice of procedure has to be carefully individualized, when feasible we favor an IPAA for most FAP patients, because of the risk of rectal cancer associated with an IRA. However, an IRA can be a consideration in specific circumstances, such as when there is mild rectal polyposis (as in attenuated FAP), or a young patient not interested in undergoing the multiple procedures that accompany an IPAA and diverting loop ileostomy. However, although we attempt to perform a diverting loop ileostomy on all IPAA procedures, this is not always feasible due to a number of anatomic factors, including body habitus. Endoscopic surveillance of the rectal segment at 6-month to 1-year intervals after the index surgery is recommended. 39 With increasing numbers of adenomas, frequency of surveillance should be increased. Although small ( 5 mm) scattered adenomas can be safely observed or removed with a biopsy forceps, polyps larger than 5 mm should be removed with a snare. However, repeated fulguration and polypectomy over many years can lead to difficulty with subsequent polypectomy, reduced rectal compliance, and difficulty identifying flat cancers within a background of scar tissue. The development of severe dysplasia or villous adenoma not amenable to endoscopic removal is an indication for proctectomy. Nonsurgical Alternatives A number of nonsteroidal anti-inflammatory drugs, including sulindac, celecoxib, rofecoxib, and the sulindac metabolite, exisulind, have been shown to reduce polyp number and size in patients with FAP However, the long-term use of chemopreventive agents for primary treatment of FAP, in lieu of surgery, is not recommended. 36,68 In a recent randomized, placebo-controlled, double-blind study of genotyped patients, sulindac did not impact on the subsequent development of colorectal polyposis. 69 Furthermore, rectal cancer has developed in patients in whom rectal polyps were effectively controlled with sulindac. 67,68,70 Finally, these medications require continued compliance, 65 and may be associated with significant adverse effects. The use of these medications, however, may reduce polyp load and facilitate endoscopic management of polyps in an ileal pouch or retained rectum in those at high risk for polyps, or those who refuse proctectomy. Nevertheless, these patients still require careful surveillance (proctoscopy or pouchoscopy) every 6 months, depending on the findings of the previous endoscopy. Long-Term Considerations From Extracolonic Manifestations Despite the reduced risk of colorectal cancer-related death after prophylactic colectomy, FAP patients are still at increased risk of mortality from both rectal cancer and other causes relative to the general population. In one study, 222 FAP patients status post-ira and under regular surveillance had a risk of death three times higher than an age- and sex-matched population. 71 The three main causes of deaths after IRA were upper gastrointestinal malignancy, progression of desmoid disease, and perioperative mortality. In another report of 354 patients without cancer at the time of diagnosis, and status post- IRA, 27 died at the time of follow-up. Causes of mortality included rectal cancer in 26%, extracolonic cancer in 30%, desmoid disease in 18%, and other causes in 26%. 72 Desmoids Desmoids are histologically benign tumors arising from fibroaponeurotic tissue. 73 They occur in 12% to 17% of patients with FAP 36,74 and, unlike those found in the general population, tend to be intra-abdominal (up to 80%) and occur after prior abdominal surgery Patients with APC mutations located between codons 1310 and 2011 appear to be at increased risk of desmoid tumors. 76 These tumors often involve the small bowel mesentery ( 50%), making complete resection difficult or impossible, and may also involve the ureters. 73 Presentation with small bowel obstruction is not uncommon Morbidity after attempted resection, which often involves removal of a variable length of small bowel, is substantial. Furthermore, the rate of recurrence is high after attempted resection, with recurrent disease often more aggressive than the initial desmoid Intra-abdominal desmoids may be more common and severe after IRA than after IPAA. 74,75,78 Desmoids involving the small bowel mesentery may preclude the formation of an IPAA secondary to foreshortening of the small bowel mesentery, especially in patients undergoing proctectomy after an initial IRA JOURNAL OF CLINICAL ONCOLOGY

6 Current Role of Surgery in Hereditary Cancer Surgery for abdominal wall desmoids should be reserved for small, well-defined tumors with a clear margin. 36 When intraabdominal desmoid tumors involve the small bowel mesentery, they should be treated according to their initial presentation and rate of growth. Tamoxifen or other antiestrogens may be considered when tumors are slow growing or mildly symptomatic. 79,80 More aggressive desmoids, in particular large intra-abdominal tumors directly invading the abdominal wall which present a formidable surgical challenge associated with high rates of morbidity and mortality, should be considered for treatment with chemotherapy. 81 The combination of vinblastine and methotrexate has achieved response in 40% to 50% of patients. 82,83 For desmoids characterized by more rapid growth, therapies used in treatment of sarcomas, such as doxorubicin and dacarbazine, may be instituted. 84,85 Radiation therapy may also be effective, but can result in substantial morbidity secondary to irradiation of the adjacent small bowel. In the case of desmoid tumors that are refractory to all medical treatment and require surgery with extensive small bowel resection, small bowel transplantation may be feasible in selected cases. 86 Because of the rarity and highly variable natural history of desmoids, efforts to establish uniformity in staging of the disease 81 will be essential for developing definitive trials and effective therapies. Upper Gastrointestinal Neoplasms Approximately 80% to 90% of individuals with FAP will develop duodenal or periampullary adenomas Of these, 36% will develop advanced polyposis and 3% to 5% will develop invasive cancer. 30,90-92 Although still relatively rare, the risk of periampullary or duodenal cancer in FAP patients is several hundred-fold greater than that in the general population. Although inconsistent, most reports indicate that mutations in exon 15 of the APC gene, particularly distal to codon 1400, are associated with the highest rate of duodenal adenomas. 93 Insofar as the most common site of upper gastrointestinal polyps is the ampullary and periampullary region, patients should begin surveillance with side viewing esophagogastroduodenoscopy and biopsy of suspicious polyps by age 25 to 30 years. The purpose of periodic endoscopy is to monitor for the development of high-grade dysplasia, rather than the removal of all visible polyps. Small, tubular adenomas without high-grade dysplasia may be biopsied and observed. However, adenomas showing high-grade dysplasia, villous changes, ulceration or size larger than 1 cm should be removed. Although endoscopic removal (including endoscopic mucosal resection and snare ampullectomy 94 ) or a transduodenal excision are options,endoscopicablationgenerallyrequiresmultiplesittings, 90,95 and recurrence is high after either procedure. 90,95,96 Endoscopic ablation is a reasonable initial approach for most patients without invasive cancer as definitive therapy for patients unfit for duodenal resection. For patients with persistent or recurrent high-grade dysplasia in papillary or duodenal adenomas, and for patients with Spigelman stage IV disease (Table 4), 30 pancreas-preserving duodenectomy or pancreaticoduodenectomy is recommended. 36 Results for duodenal resection in patients with premalignant lesions are encouraging, with good local control and low morbidity. 90,97-99 Of note, with the exception of one study suggesting a clinical response with celecoxib, 100 chemoprevention has not been proven effective in the management of duodenal adenomas. HNPCC The diagnosis of HNPCC (or Lynch Syndrome ) is much more challenging than FAP, as it requires ascertainment of an accurate and Table 4. The Spigelman Classification for Staging Duodenal Polyposis Characteristic Grade of Duodenal Disease (points) No. of polyps Polyp size, mm Histologic type Tubular Tubulovillous Villous Dysplasia Mild Moderate Severe NOTE. Adapted from Guillem et al with permission. 30 Stage 0, 0 points; stage 1, 1-4 points; stage II, 5-6 point, stage III, 7-8 points; stage IV, 9-12 points. detailed family history, and an awareness of the syndrome by the clinician. However, it is important to recognize that although a family history of colorectal cancer is helpful when present, its absence does not exclude HNPCC. The Amsterdam criteria (I and II) require that there be three relatives (one a first-degree relative of the other two) with an HNPCC-related cancer (colorectal, endometrial, small bowel, ureter, or renal pelvis), involving two or more successive generations, with at least one with colorectal cancer diagnosed at younger than 50 years of age. 101,102 Finally, FAP should be excluded. It is important to keep in mind that not all individuals with HNPCC will meet these criteria; thus, the clinician has to have a high index of clinical suspicion for HNPCC in order to diagnose it. It is equally important to recognize that only approximately 60% of families who meet the Amsterdam criteria have a hereditary abnormality in an MMR gene. 31 The term familial colorectal cancer type X has been suggested in order to distinguish families who fulfill Amsterdam criteria but do not have evidence of a DNA MMR defect, because relatives in these families appear to have a lower incidence of colorectal cancer relative to those belonging to a family in whom an MMR mutation has been detected. 103 The Bethesda criteria, which are more inclusive, were established 104 and revised 105 in order to identify individuals and families in which HNPCC is suspected and for whom MSI testing is indicated (Table 5 ). Overall, colorectal cancer occurs in 78% to 80% of MMR mutation-positive patients, at a mean age of 46 years. 30,37,106,107 Endometrial cancer occurs in 43%, gastric cancer in 19%, urinary tract cancer in 18%, and ovarian cancer in 9% of affected individuals. 108 Patients with colorectal cancer belonging to known HNPCC kindreds, or pedigrees suspicious for HNPCC, should be offered screening of their tumor by MSI analysis, with consideration for subsequent immunohistochemical (IHC) evaluation for loss of MMR expression. Recently, however, it has been shown that screening with IHC for loss of MMR proteins expression is as effective as screening with the more complex MSI analysis. 109 Although perhaps not as commonly noted as in BRCA testing for familial breast and ovarian cancer, MMR testing for HNPCC can disclose variants of uncertain significance or cases of incomplete or variable penetrance which mandate careful counseling. MSI evaluation will be positive (MSI-high) in more than 80% of patients belonging to families that meet Amsterdam criteria. Patients with MSI-high tumors should undergo testing for germline MMR mutations(hmsh2, hmlh1, and MSH6 are currently commercially available). IHC staining for hmsh2 and hmlh1 protein expression in colorectal tumors have been shown to be highly sensitive and specific in screening for MMR gene defects. 110 Recently, PMS2 staining has been shown to detect additional cases with hmlh

7 Guillem et al Table 5. The Revised Bethesda Guidelines for Testing Colorectal Tumors for MSI 105 Tumors From Individuals Should Be Tested for MSI in the Following Situations Colorectal cancer diagnosed in a patient who is younger than 50 years of age Presence of synchronous, metachronous colorectal, or other HNPCCassociated tumors, regardless of age Colorectal cancer with the MSI-H histology diagnosed in a patient who is younger than 60 years of age Colorectal cancer diagnosed in one or more first-degree relatives with an HNPCC-related tumor, with one of the cancers being diagnosed at younger than age 50 years Colorectal cancer diagnosed in two or more first-or second-degree relatives with HNPCC-related tumors, regardless of age Abbreviations: MSI, microsatellite instability; MSI-H, microsatellite instabilityhigh; HNPCC, hereditary nonpolyposis colorectal cancer. HNPCC-related tumors include colorectal, endometrial, stomach, ovarian, pancreas, ureter and renal pelvis, biliary tract, and brain (usually glioblastoma as seen in Turcot syndrome) tumors, sebaceous gland adenomas and keratoacanthomas in Muir-Torre syndrome, and carcinoma of the small bowel. MSI-H in tumors refers to changes in two or more of the five National Cancer Institute recommended panels of MS markers. Presence of tumor infiltrating lymphocytes, Crohn s-like lymphocytic reaction, mucinous/signet ring differentiation, or medullary growth pattern. There was no consensus among the workshop participants on whether to include the age criteria in guideline 3 above; participants voted to keep younger than 60 years of age in the guidelines. germline mutations. 111 In addition, an abnormal IHC has been shown to have a 100% predictive value for MSI-H. 112 In families in which tumor tissue is not available, initial germline testing may be considered. As in FAP, a mutation in an affected individual must be established for testing in at-risk individuals to be conclusive. 39 Recommended surveillance for HNPCC includes colonoscopy every1to2years,beginningat20to25years,andannuallyafterage40. Given the increasing evidence for an accelerated adenoma-carcinoma sequence in HNPCC, annual colonoscopy should be strongly considered. 31 For females, consideration of annual endometrial aspiration is also recommended, starting at 30 to 35 years. Annual esophagogastroduodenoscopy is recommended for patients belonging to kindreds with a history of gastric cancer. Finally, ultrasonography and urine cytology every 1 to 2 years should be considered to screen for urinary tract malignancy in pedigrees with a predilection for genitourinary tumor. 113 It must be noted, however, that in HNPCC the colon and rectum are the only target organs where surveillance has been proven effective for reducing cancer. Although development of colorectal cancer in HNPCC is not a certainty,the80%lifetimerisk, 30 45%rateofmetachronouscolorectal neoplasms, 106 and the possible accelerated adenoma to carcinoma sequence 31 mandates consideration of prophylactic surgical options. Patients with HNPCC as defined by genotype or Amsterdam criteria, with a colon cancer or more than one advanced adenoma, should be offered the options of prophylactic total colectomy with IRA or segmental colectomy with annual lower endoscopy. 46, In patients undergoing IRA, the goal should be normal rectal and sphincter function. Although the risk of metachronous colon cancers is higher after partial colectomy versus total colectomy with IRA, intensive colonoscopic surveillance and polypectomy may minimize the risk of future cancers in the remaining colon. 101,113 Careful surveillance is also necessary after total colectomy and IRA because the risk of cancer in the retained rectum is approximately 12% at 10 to 12 years. 118,119 Although there is no trial demonstrating an improved survival for HNPCC patients undergoing a total colectomy and IRA rather than a segmental colectomy, mathematical models suggest benefit for total colectomy and IRA, especially for younger individuals with early-stage cancers. 120 Mismatch mutation positive patients with a healthy colon may also be offered prophylactic colectomy in highly selected situations. 121 One rationale for this approach is the similarity of lifetime cancer risk between patients with APC and MMR gene mutations, and the fact that total abdominal colectomy with IRA produces less functional disturbance than the prophylactic procedure recommended for FAP (total proctocolectomy with IPAA). 121 However, an alternate strategy in these individuals is surveillance by colonoscopy at 1- to 3-year intervals, which is cost effective 117,122 and greatly reduces the rate of colorectal cancer development and overall mortality. 107, There is a risk of colorectal cancer development in the interval between colonoscopies, 123,126 but when the interval is less than 2 years these are often early-stage, curable tumors. 107,123 A decision analysis model suggests that prophylactic subtotal colectomy at age 25 may offer a survival benefit of 1.8 years compared with surveillance colonoscopy. The benefit of prophylactic colectomy decreases when surgery is delayed until later in life, and is negligible when performed at the time of cancer development. 124 However, when quality of life was considered, surveillance provided the greatest benefit in quality-adjusted life years. 119 Based on this evidence, prophylactic surgery is clearly indicated only in those patients for whom colonoscopic surveillance is not technically possible or who refuse to undergo regular surveillance. 125 Thus, the decision between prophylactic surgery and surveillance for gene positive, unaffected patients is based on many factors, including penetrance of disease in a particular family, early age of onset in affected family members, functional and quality of life considerations, and likelihood of compliance with surveillance. HNPCC patients with an index rectal cancer should be offered the options of total proctocolectomy with IPAA or anterior proctosigmoidectomy with primary reconstruction. 36,116 The rationale for total proctocolectomy is the 17% to 45% rate of metachronous colon cancer in the remaining colon after an index rectal cancer in HNPCC patients. 119,126,127 The decision between the two procedures depends, in part, on the patient s willingness to undergo intensive surveillance of the retained proximal colon, as well as issues of bowel function, which differs after each of these procedures. Management of Extracolonic Cancers Management of extracolonic cancers in HNPCC patients is less well defined. Prophylactic total abdominal hysterectomy should be offered to female patients whose childbearing is complete or to women undergoing abdominal surgery for other conditions, 46 especially when there is endometrial cancer in the family. (There is evidence of higher risks for endometrial cancer in MSH2 and MSH6 mutation carriers.) This recommendation is based on the high rate of endometrial cancer in mutation positive individuals (43%), 108 and the lack of efficacy of screening demonstrated by some studies. 128 Oophorectomy should also be performed because of the high incidence of ovarian cancer in HNPCC (9%), 108 and the frequent coexistence of endometrial and ovarian cancer. 129 Optimal timing of prophylactic total abdominal hysterectomy is unclear. However, endometrial 4648 JOURNAL OF CLINICAL ONCOLOGY

8 Current Role of Surgery in Hereditary Cancer cancer has been reported in HNPCC patients younger than age 35. At present, it seems reasonable to begin surveillance at age 25 and delay prophylactic surgery until childbearing is complete. 36 These issues are discussed in greater detail in Hereditary Ovarian and Endometrial Cancers. HEREDITARY OVARIAN AND ENDOMETRIAL CANCERS Hereditary Ovarian Cancer Germline mutations in the BRCA1 and BRCA2 genes strongly predispose women to early onset breast and ovarian cancers. Fallopian tube cancers are rarer, but also appear to have an increased incidence in these families, and in many instances it is not possible to determine with certainty whether the cancer arose in the ovary or tube. Although only a minority of young women with early onset breast cancer and a strong family history are carriers, 130 it appears that most familial ovarian cancer is attributable to BRCA1 and BRCA2 which account for approximately 10% of all invasive epithelial ovarian cancers The risk of ovarian cancer is also increased in HNPCC families, but this accounts for only approximately 1% of all ovarian cancers. 131 The lifetime risk of ovarian cancer increases from a baseline of 1.5% to approximately 5% to 10% in HNPCC carriers, 10% to 20% inbrca2 carriers, and 20% to 40% in BRCA1 carriers Highly penetrant germline BRCA mutations are rare, however, and in most populations are carried by fewer than one in 500 individuals. One notable exception is the Ashkenazi Jewish population, with a carrier frequency of one in Ovarian cancer is relatively uncommon, but has a high mortality rate because the majority of cases present at a late stage and effective early detection methods do not exist. As a result, most women present with extensive abdominal carcinomatosis and therapy is generally aimed at achieving remission rather than cure. The strategy of performing risk-reducing salpingo-oophorectomy (RRSO) in BRCA mutation carriers has the potential to significantly decrease ovarian cancer mortality in these individuals. In the past, RRSO was performed in women with a strong family history of ovarian cancer, whereas currently the decision to proceed with RRSO is based primarily on the results of BRCA mutational analysis. Genetic Testing As It Relates to Decisions Regarding Prophylactic Oophorectomy Although BRCA1 and BRCA2 play a role in the repair of doublestranded DNA breaks, the molecular basis for the increased risk of breast, ovarian, and fallopian tube cancers but not other cancer types in mutation carriers is unclear. At least twice as many BRCA1 mutations as BRCA2 mutations are found in high-risk families with a history of ovarian cancer. 138 Most deleterious BRCA mutations encode truncated protein products, but missense mutations that alter a single amino acid have been found to segregate with breast or ovarian cancer in some families. 138,139 In a significant fraction of high-risk families, BRCA testing reveals sequence variants of uncertain significance or no detectable alterations, and these results represent a counseling dilemma. Incomplete or variable penetrance of disease is another issue that confounds decision making for women considering prophylactic oophorectomy. There is evidence that certain types of BRCA mutations may predispose more strongly to ovarian cancer, 140,141 buttherelationshipisnotstrongenoughtoaffectcounseling of patients. Surgeons performing RRSO in high-risk women should be knowledgeable regarding the risks and benefits of prophylactic surgery and prepared to discuss these issues with patients. Genetic counselors should also play a significant role in ongoing patient management. Because it has been feared that misuse of genetic information could have devastating consequences, including difficulty in securing employment and life, health, or disability insurance, clinicians were initially hesitant to record genetic testing results in the medical record. However, since BRCA testing is now widely accepted and insurance companies generally cover the costs, patients should be encouraged to allow test results to be recorded in the medical record, as they form the basis for the decision to perform prophylactic surgery. Because of ovarian cancer s high mortality rate, RRSO should be strongly recommended to all women who carry germline BRCA1 or BRCA2 mutations. Fortunately, the incidence of ovarian cancer in mutation carriers does not begin to rise dramatically until the patients are in their late 30s. 142 This allows women time to bear children before considering RRSO as they approach the end of their reproductive life span. Although oophorectomy is a major surgical procedure, surveys show that approximately three quarters of high-risk women undergoing genetic testing would strongly consider RRSO. 143 The ovaries are internal organs, and most women experience only modest feelings of altered body image and self-esteem after oophorectomy. In contrast, because cure rates for breast cancer are high and the cosmetic and emotional consequences of prophylactic mastectomy are more significant, surgical prophylaxis is accepted by only approximately one third of women. Insurance companies have almost always paid for RRSO in proven mutation carriers. 144 High-risk women lacking BRCA mutations, or those with sequence alterations of uncertain significance, may encounter greater reimbursement obstacles. For younger women who wish to maintain fertility, periodic screening with CA125 and transvaginal sonography, although of unproven benefit, is often advised. The National Cancer Institute (Bethesda, MD) Cancer Genetics Network is presently performing a pilot study of these modalities in high-risk women. RRSO RRSO is widely viewed as the most effective means currently available of decreasing ovarian cancer mortality in BRCA mutation carriers. The risks and benefits of surgical prophylaxis for hereditary ovarian cancer are summarized in Table 6. First, there is strong evidence that this approach significantly decreases ovarian cancer mor- Table 6. Advantages and Disadvantages of Risk-Reducing Salpingo-Oophorectomy Advantage Strikingly decreases ovarian and fallopian tube cancer incidence and mortality Can be delayed to allow completion of childbearing Laparoscopic approach possible in most cases Impact on body image generally acceptable Estrogen replacement can prevent consequences of surgical menopause Decreases breast cancer risk Disadvantage Cost Potential operative morbidity and mortality Surgical menopause in premenopausal patients who elect not to take hormone replacement